Transient stability assessment using artificial neural network

This paper proposes a novel method of transient stability assessment by adaptive pattern recognition which makes use of an artificial neural network (ANN). A multilayer feed forward network has been applied to a sample nine bus three machine system. A test set pattern chosen outside the region of the learning set patterns giving very close to desirable output (error below 5%) validates the generalising and extrapolating capability of the network after learning. The method has been successfully applied to determine which particular machine goes out of step.